Comment 15 for bug 1579190

Doug, you added the tag 'kernel-fixed-upstream' but I don't think we have hard evidence of that at this point. Was your tagging based only on Viktor's reports that it hasn't happened (yet) on a 4.6 kernel, or is there more information elsewhere?

Viktor, your result with 4.6 is promising! However, since this problem is known to "come and go" it is also possible that code has just moved around such that its just not manifesting in your current 4.6 kernel build you currently (but could reappear in some future build of 4.6).

I do see one potentially interesting fix in 4.6 (50220de HID: core: prevent out-of-bound reading). I'm going to produce a test kernel with just that fix backported on top of Xenial 4.4.0-22.40 (which I've seen manifest the problem), and will post that here for further testing.